1. Building the project team
By now, we may be thinking, “who is going to build this thing?”
2. The project team
The make-up of the project team can vary depending on the use case, what is built or bought, and the size of the company. The roles typically cover three broad areas: machine learning, software engineering, and business.
3. The machine learning roles
Machine learning includes roles such as data analyst and a machine learning, or ML, scientist. A data analyst will focus on exploring the data and transforming it for the AI. They may support monitoring and solution evaluation as well.
ML scientists are vital for developing the AI model, whether building or buying. They will focus on selecting the appropriate algorithm and ensuring it learns patterns effectively. In the scenario that an AI solution was bought to be integrated into business processes, they will be important as a subject matter expert consultant for the project.
4. The software engineering team
Software Engineering includes roles such as solution architect, data engineer, product engineer, UX designer, and operations (development and ML). This is a big part of the project team.
5. The software engineering team
The solution architect will be the point person for helping design the architecture of the system. This includes cloud services, compute requirements, and other technology necessary for a working POC. In some cases, a project manager can fill in this role, ie, a Technical PM.
Data engineers will be important for building the necessary data pipelines to move, aggregate, transform, and store data at scale. Sometimes, especially for a small POC, the data analyst or ML scientist may be able to do this themselves. However, make sure to account for this added effort on their part.
6. The software engineering team
Product engineers, sometimes called software engineers, will build the actual product, or feature within a product, in which the AI model will live. Ultimately, one or more are needed with the skills to quickly and efficiently write code to create a functioning product back-end and front-end.
User Experience, or UX, designers will be important for creating an experience that will elicit the desired interactions and responses from the end-users with the AI solution. They help ensure an intuitive user interface which in turn helps toward the goal of usability.
7. The software engineering team
Operations roles will be focused on two areas - MLOps engineer for the AI model, and DevOps engineer for the product.
The former will support automating and productionalizing the model. They will help build monitoring and alert systems to catch changes in performance of the AI.
The latter will support the solution development lifecycle to enable continuous updates. These roles may not be 100% necessary for a POC, but should definitely be included in scaling conversations.
8. The business roles
Business includes roles such as project manager and business expert. A project manager will be essential to ensuring the POC has clearly defined requirements, building a timeline, and driving clarity.
The business expert will bring the industry and end-user expertise. This role can be anyone from the VP of Marketing to a program manager for customer support. They will be more of a consultant but important in use case identification and evaluating impact.
9. Third-party support
Not every company will have these roles readily available. For some POCs, it may be more effective to hire an outside development partner. These agencies may even be able to support change management, upskilling, or have existing AI solutions to enable faster development times. They may even have more data to utilize!
10. Let's practice!
These are the general roles to include in the conversation. Now, let's test your knowledge of them!